Skip to content
GuideBeginner

Why the Number on the Screen Is Not Always the Number You Can Trade

Overlapping cardboard tags stamped MID and ASK.
Displayed price is often a midpoint. Buyers pay the ask. Sloppy Stakes editorial illustration

The displayed price is often a midpoint. Buyers pay the ask. Taker fees apply on many categories. A Yes+No sum that looks like $1 can still not be a lock.

By Jordan NabigonPublished 8 min read

What you'll be able to do

  • Distinguish displayed midpoint, last trade, bid, and ask on the documented venue.
  • State the published taker-fee formula and category table as documentation, not as a live invoice.
  • Leave the current fee on any named live market UNKNOWN unless that market's feeSchedule is separately read.
  • Explain why a Yes+No sum that looks like $1 can still fail as a lock: display, spread, size, two orders, taker fees.
  • Refuse the readings this page does not support: buy or sell the gap; the screenshot is the fill; fee tables are your bill.

Before you start

  • what-is-a-prediction-market
  • what-a-65-percent-price-means

What this guide won't do

  • No live prices. The $0.34/$0.40 display pair, the $0.60+$0.40 mint pair, and the 100-share fee tables are documentation examples retrieved 2026-08-15T03:10Z.
  • Display, fee, and rebate rules cited from one venue's public docs; they are not universal and are dated by retrieval.
  • Current fee on any named live market is UNKNOWN. Per-market feesEnabled/feeSchedule was not fetched live.
  • Companion page owns the $1 identity, δ, and historical under/over findings. This page does not census 2026 gaps.
  • No trading advice. Wallet profitability and named extraction cases are separate pieces.

The number on the screen is often a midpoint. It is not always the number a buy or a sell would fill at. On fee-on books, a taker fee sits on top of that fill. Two displayed prices that add up to about a dollar can still fail as a lock: they may be the wrong two numbers, they may not be executable in size, they are two orders rather than one click, and on many categories they are two fee-on fills.

What this page does not claim. It does not tell you to buy or sell. It does not claim a Yes+No sum near $1 is money on the table. It does not treat a documentation fee table as your invoice, or guess the current fee on a named live market. The $0.34-and-$0.40 display pair, the $0.60-and-$0.40 mint pair, and the 100-share fee tables below are documentation examples published by one venue, retrieved 2026-08-15T03:10Z, not live quotes. Why complementary claims should add to about a dollar — split, merge, redeem, δ — is a companion guide. This page is about the numbers you can see and the costs the docs put between those numbers and a trade.

The displayed price is usually a midpoint

On the venue that published the prices-and-orderbook page retrieved 2026-08-15T03:10Z, the displayed price is the midpoint of the best bid and the best ask. If the spread is wider than $0.10, the last trade is shown instead.

Their own example: best bid $0.34, best ask $0.40. Displayed price = ($0.34 + $0.40) / 2 = $0.37. A buyer pays the ask ($0.40). A seller receives the bid ($0.34). You will not necessarily trade at $0.37.

That $0.34 / $0.40 / $0.37 trio is a documentation example, not a live book.

The same venue's market-data docs draw the same line in API form. A midpoint call is the average of best bid and best ask. A buy-side "price" call returns the lowest ask. A sell-side call returns the highest bid. Those are different objects. A headline that says "37%" may be pointing at the midpoint. A fill is pointing at bid or ask.

If the interface has fallen through to last trade, you are looking at a print from some earlier moment. The last-trade API documents a further trap: when a token has never traded and the book is empty, the placeholder price is 0.5 with no side. Two placeholders add to a dollar and still describe no trade.

Spread is the gap you cross

The spread is the gap between the highest bid and the lowest ask. The docs describe tighter spreads as more liquid. They also say large orders may move the price significantly, and that you should check order-book depth — size resting at each level — before assuming the best quote is the whole fill.

Orders can fill in pieces. A "market order" on this venue is still a limit order, priced to cross the book. Minimum order size and tick size are per market. The market-details page's example of a 1¢ tick and a minimum size of 5 is documentation, not a universal rule. The page says to read the active values from the market.

None of that is a recommendation to take or to rest. It is why the number on the button and the number in a fill report are allowed to differ.

Who pays a trading fee

As of the fees page retrieved 2026-08-15T03:10Z, takers on certain markets pay a fee. Makers are never charged a trading fee. Fees are set by the protocol and applied at match time; they are not a field you put in the order.

The published formula is:

fee = C × feeRate × p × (1 − p)

C is the number of shares. p is the price of those shares. The USDC fee is symmetric around 50%: a trade at 30¢ incurs the same dollar fee as a trade at 70¢. The dollar fee peaks at a 50¢ price.

Documented taker feeRate on that page: Crypto 0.07; Sports, Economics, Culture, Weather, and Other / General 0.05; Finance, Politics, Mentions, and Tech 0.04; Geopolitics 0. Geopolitical and world-events markets are described as fee-free. The page also says there are no Polymarket fees to deposit or withdraw USDC; intermediaries may charge their own. Those intermediary amounts are UNKNOWN here.

The same page's 100-share tables are documentation examples, not a live invoice. At 50¢ they peak at $1.75 (Crypto), $1.25 (0.05-rate categories), and $1.00 (0.04-rate categories) per 100 shares. Fees round to five decimal places. The smallest fee charged is 0.00001 USDC; smaller amounts round to zero.

Qin and Yang (arXiv:2606.04217v1, HTML retrieved 2026-08-15T03:10Z) date the 2026 fee reform as staggered: Crypto in January 2026, Sports (selected leagues) in February 2026, other categories in March 2026. On the v1 tape they study, fee revenue was zero before January 2026. That timeline is not a live fee for a market you have open tonight.

The fee on a named live market is not in this table

The fees page tells integrations to read fee parameters from Market Details. That page exposes feesEnabled and a feeSchedule — rate, exponent, taker-only flag, rebate rate — on the market object. Its documentation JSON uses rate 0.04, exponent 1, takerOnly true, rebateRate 0.25. That JSON is not an invoice.

This packet did not fetch a live market object. The current fee on any named live market is UNKNOWN. Category of a named market is UNKNOWN until that market is read. Whether any live book uses an exponent other than 1 is UNKNOWN: the fees-page formula writes p × (1 − p) with no extra exponent; the market-details example sets exponent to 1.

Taker fees fund a maker-rebate program (daily pUSD, $1 minimum payout, category percentages on the maker-rebates page as of retrieval). A separate taker-rebate program is documented as going live 28 May 2026, with tiered percentages that the page says can change. Net fee after rebate for any named person is UNKNOWN. This page does not tell you to chase rebates.

Why a Yes+No sum that looks like $1 can still not be a lock

The companion guide is the place for why complementary claims are two sides of one dollar of collateral. Official docs on this venue (retrieved 2026-08-15T03:10Z) still matter here for one limited fact: a buy-Yes at $0.60 can match a buy-No at $0.40 because those bids add to $1.00, and a dollar is then converted into one Yes and one No. That $0.60 / $0.40 pair is a documentation example, not a live fill. It is a description of a mint path when complementary bids actually meet.

A screenshot is usually not that object.

If the two numbers you added are displayed mids, you added two averages. Each buyer of a leg pays the ask. Two half-spreads sit between mid+mid and ask+ask. On the docs' own Yes example, that half-spread is three cents on one leg ($0.37 displayed, $0.40 to buy). This page will not invent a live No book to complete the pair. The arithmetic is enough: a mid-sum near $1 does not mean the executable pair costs $1.

If one or both numbers are last trades, you may be adding prints from different times — or the 0.5 placeholder the last-trade API returns when nothing has traded.

If the two numbers are executable asks, you still have size. The best ask may be thinner than the screenshot implies. Orders can fill in part. The two legs are two orders. Settlement of a single matched trade is atomic; the docs do not describe Yes and No as one bundled click. Some markets delay a marketable taker order (the order-lifecycle page: 250 milliseconds on selected crypto and finance up/down markets; a sports delay on configured games). During that delay the order cannot be cancelled.

On fee-on categories, each taker fill can pay C × feeRate × p × (1 − p). Two legs can mean two fees. At the documentation peak (100 shares at 50¢) that is $1.00, $1.25, or $1.75 per leg depending on the published rate group — a documentation table, not your invoice, and not a claim about a named book. A gap that is smaller than two half-spreads plus two taker fees plus the chance one leg misses is not a lock. Whether any given 2026 book still shows an executable complementary gap after those costs is UNKNOWN. This page does not claim the historical gaps in the companion's papers are still there, and it does not tell you to take either side.

What to do with the number on the screen

Read which object it is. Mid, last, bid, and ask are different. If the spread is wider than ten cents, the documented UI may have switched to last trade. If you cannot see bid, ask, size, and whether fees are on, you do not have an executable price.

Read whether the book is in a fee-on category only from that market's own parameters — not from this page's table applied to a name. The current fee on a named live market is UNKNOWN here.

If Yes and No appear to add to a dollar, send that screenshot to the companion question: are these complementary claims, and which two numbers were added? Then bring it back here: even a true complementary pair can fail as a lock once display, spread, size, two orders, and taker fees are in the picture.

Wallet profitability, and named wallets a paper measured against historical gaps, are separate pieces. They are not evidence that a displayed dollar is yours.

Editorial review

Reviewed by Jordan Nabigon on 15 August 2026.

Sources

  • Prices & Orderbook primaryPolymarket · Official explanation of displayed probabilities, bids, asks, midpoint logic, and order-book pricing.
  • Fees UnverifiedPolymarket · retrieved 15 August 2026
  • What Is a Prediction Market?

    A prediction market is a market in contracts that pay a fixed amount if a defined event happens. The price is what people will pay for that payoff right now — useful as a probability estimate, but not a fact about the world.

  • What one academic paper documented about complementary-price extraction on Polymarket

    Saguillo et al. (AFT 2025) classified complementary-price, NegRisk, and combinatorial extraction on markets that resolved 1 April 2024 to 1 April 2025 and reported $39,587,585.02. That is a pre-fee, dated measurement — not “bots made $40 million,” and not a claim the same gaps are still there in 2026.

  • The Briefing

    What moved in prediction markets, in short.